A method and system for identifying and preventing and controlling the risk of evaluation of residual risk based on data mining

Abstract

The application provides a kind of based on data mining's risk identification and prevention and control analysis method and system of evaluation of residual, it is related to data processing technical field, method includes: acquisition each monitoring signal corresponding multi-source function monitoring data;Pretreatment obtains standardized multi-source function monitoring data;The change amplitude, complexity and stability characteristics of monitoring signal are extracted, and the function degradation trend index is quantified;Extract the mutation characteristics of monitoring signal, quantify short-time fluctuation and energy anomaly, obtain function burst response intensity value;Identify potential function degradation trend and phase mutation event;Extract the synergistic change characteristics of monitoring signal, analyze the multi-dimensional function coupling relationship;Build basic quantization parameter, combine multi-dimensional function coupling relationship, quantify individual multi-dimensional function degradation degree;Dynamic identification and hierarchical management are carried out;Continuously monitor and visualize function degradation trend index, function burst response intensity value and individual multi-dimensional function degradation degree, dynamically display function degradation trend and burst event.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for disability assessment risk identification and prevention based on data mining. Background Technology

[0002] The field of disability is an important component of the social public service system, relating to people's livelihood and social equity. With the accelerating aging of the population and the rising prevalence of chronic diseases, the number of people with functional impairments due to injury or illness continues to expand, leading to a growing demand for rehabilitation, assessment, and management of people with disabilities. National and social institutions at all levels are continuously strengthening the dynamic monitoring and scientific assessment of the functional status, rehabilitation progress, and self-care abilities of people with disabilities to support precise services, categorized guidance, and the rational allocation of social security resources.

[0003] Currently, in the field of continuous monitoring of the functional status of people with disabilities, there exists a long-term data collection and evaluation method based on wearable devices. This method uses sensors such as triaxial accelerometers and heart rate belts to periodically (e.g., daily or weekly) collect raw data such as the user's cadence and heart rate, and calculates basic statistics such as the average and standard deviation over a period of time (e.g., weekly) as a basis for assessing functional status. For example, comparing the average cadence of this month with that of last month can help determine if motor ability has declined.

[0004] However, existing technologies lack the ability to perform high-sensitivity, high-resolution dynamic feature mining on multi-source functional monitoring signals. They cannot simultaneously extract and quantify in-depth indicators that comprehensively reflect both the "strength of long-term gradual trends" and the "strength of short-term sudden responses" from continuous data streams. This results in a significant lag in identifying potential functional degradation risks and a weak ability to capture phased functional abrupt events. Fundamentally, this limits the possibility of the system achieving early, accurate, and categorized early warnings. Summary of the Invention

[0005] To address the lack of high-sensitivity, high-resolution dynamic feature mining capabilities in existing technologies for multi-source functional monitoring signals, and the inability to simultaneously extract and quantify in-depth indicators that comprehensively reflect both "long-term gradual trend intensity" and "short-term sudden response intensity" from continuous data streams, resulting in significant delays in identifying potential functional degradation risks and weak ability to capture staged functional abrupt events, this invention fundamentally limits the possibility of achieving early, accurate, and categorized early warning systems. Therefore, this invention provides a data mining-based method and system for disability assessment risk identification and prevention analysis.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] The first aspect of this invention provides a data mining-based method for disability assessment risk identification and prevention analysis, comprising:

[0008] S1: Collect multi-source functional monitoring data corresponding to each monitoring signal;

[0009] S2: Filter, resample and normalize the multi-source functional monitoring data to obtain standardized multi-source functional monitoring data;

[0010] S3: Based on standardized multi-source functional monitoring data, extract the variation amplitude, complexity and stability characteristics of each monitoring signal, and quantify the functional degradation trend indicators;

[0011] S4: Based on standardized multi-source functional monitoring data, the sudden change characteristics of the monitoring signals are extracted, short-term fluctuations and energy anomalies are quantified, and the functional burst response intensity value is evaluated.

[0012] S5: Identify potential functional degradation trends and phased abrupt events based on functional degradation trend indicators and functional sudden response intensity values;

[0013] S6: Based on the potential functional degradation trend and stage-specific abrupt events, extract the synergistic change characteristics of the monitoring signals and analyze the multidimensional functional coupling relationship;

[0014] S7: Based on the characteristics of coordinated change, the trend of functional degradation and the intensity of sudden functional response, construct basic quantitative parameters for risk assessment, and combine them with multidimensional functional coupling relationships to quantify the degree of multidimensional functional degradation of individuals;

[0015] S8: Based on the degree of multidimensional functional decline of individuals, conduct dynamic identification and hierarchical management of disability risk;

[0016] S9: Based on the grading results, by continuously monitoring and visualizing functional degradation trend indicators, functional emergency response intensity values ​​and individual multidimensional functional degradation degree, the functional degradation trend and emergency events are dynamically displayed; and based on historical data, the threshold is optimized to achieve closed-loop self-learning of prevention and control strategies.

[0017] A second aspect of this invention provides a data mining-based disability assessment risk identification and prevention analysis system, comprising:

[0018] processor.

[0019] The memory stores computer-readable instructions, which, when executed by the processor, implement the data mining-based disability risk identification and prevention analysis method as described in the first aspect.

[0020] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the data mining-based disability assessment risk identification and prevention analysis method of the first aspect.

[0021] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0022] This invention, by simultaneously extracting the variation amplitude, complexity, and stability characteristics of multi-source functional monitoring signals, fully mines the deep degradation trend information contained in the data stream. Furthermore, it introduces a functional burst response intensity quantification method to dynamically identify short-term fluctuations and energy anomalies, improving the ability to characterize the dynamic evolution of functional states. This allows for the simultaneous capture of slow, continuous functional degradation trends and sudden, drastic, phased functional mutation events, enhancing the timeliness and accuracy of disability risk identification and reducing early warning delays and missed event reports. It also increases the system's ability to achieve early, accurate, and categorized early warnings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a data mining-based method for disability risk identification and prevention analysis, as provided in an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram illustrating the trend of the functional change rate gain value, provided for an embodiment of the present invention.

[0026] Figure 3 This is a radar diagram illustrating the risk ratio of various monitoring signals, provided as an embodiment of the present invention.

[0027] Figure 4 This is a schematic diagram of the structure of a data mining-based disability assessment risk identification and prevention analysis system provided in an embodiment of the present invention. Detailed Implementation

[0028] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0029] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0030] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0031] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0032] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0033] Reference manual attached Figure 1 The diagram shows a flowchart of a data mining-based method for disability assessment risk identification and prevention analysis provided by an embodiment of the present invention.

[0034] This invention provides a data mining-based method for disability assessment risk identification and prevention analysis. This method can be implemented using a data mining-based disability assessment risk identification and prevention analysis device, which can be a terminal or a server. The processing flow of the data mining-based disability assessment risk identification and prevention analysis method may include the following steps:

[0035] S1: Collect multi-source functional monitoring data corresponding to each monitoring signal.

[0036] In this context, the monitoring signal refers to the specific physiological or functional dimension that is observed and measured.

[0037] Among them, multi-source functional monitoring data refers to the sequence of raw measurement values ​​that change over time and are actually collected by sensors from each of the above monitoring signals.

[0038] Specifically, multi-source functional monitoring data corresponding to each monitoring signal are collected in real time. The multi-source functional monitoring data includes: acceleration, grip force, electromyography (RMS), heart rate, and skin conductance. Step frequency and stride length are calculated simultaneously based on the acceleration signal through peak detection and time integration. The grip force recovery rate is obtained by calculating the slope of the grip force rise and recovery phases according to the grip force sequence. The heart rate variability is calculated by time domain statistical analysis of the RR interval sequence of the heart rate.

[0039] S2: Perform filtering, resampling, and normalization preprocessing on the multi-source functional monitoring data to obtain standardized multi-source functional monitoring data.

[0040] Filtering refers to a signal processing technique used to remove or reduce unwanted interference components from raw data while preserving useful signal components as much as possible.

[0041] Resampling refers to the process of changing the sampling rate of a data sequence. It typically includes interpolation and decimation.

[0042] Normalization refers to a data scaling technique that maps data from different sources and with different dimensions to the same numerical range or distribution, eliminating analytical bias caused by differences in dimensions and absolute values.

[0043] In one possible implementation, S2 specifically includes:

[0044] By using a unified timestamp, multi-source functional monitoring data is aligned and resampled.

[0045] The unified timestamp refers to marking all data streams from different sensors and collected at different start times with a common, accurate, and synchronized reference time.

[0046] Random noise suppression is achieved in acceleration and electromyography (RMS) data from multi-source functional monitoring by combining bandpass filtering and moving average.

[0047] Bandpass filtering refers to a filter that only allows components within a specific frequency range of a signal to pass through, while attenuating frequency components outside the passband.

[0048] The moving average is a simple time-domain filtering method. For each point in the data sequence, the value of that point is replaced by the average of several of its neighboring points.

[0049] In this context, RMS (Root Mean Square) refers to the root mean square value of the electromyographic (EMG) signal. The original surface EMG signal is an alternating voltage signal with drastic fluctuations. The RMS value is the result of squaring, averaging, and then taking the square root of the signal amplitude over a short time window.

[0050] The Kalman filter algorithm was used to remove drift trends from heart rate, heart rate variability, and skin conductance in multi-source functional monitoring data.

[0051] The Kalman filter algorithm is an optimal estimation algorithm applicable to dynamically linear systems. It recursively predicts and corrects based on the system's state at the previous time step and the current observation, thereby obtaining the optimal estimate of the current state.

[0052] Heart rate variability refers to the minute variations in the time interval between consecutive heartbeats. It reflects the regulatory function of the cardiac autonomic nervous system and is an important indicator for assessing physiological stress, fatigue, and health status.

[0053] Skin conductivity refers to the electrical conductivity of the skin surface. It is controlled by the activity of sweat glands, which are directly innervated by the sympathetic nervous system in the autonomic nervous system.

[0054] Z-score normalization and sliding window smoothing are performed on all multi-source functional monitoring data to obtain standardized multi-source functional monitoring data.

[0055] Specifically, a disability assessment risk data mining database is established, and the raw and pre-processed multi-source functional monitoring data are written into the disability assessment risk data mining database.

[0056] In this embodiment of the invention, targeted filtering, synchronization alignment, and scale unification are performed on multi-source heterogeneous raw physiological monitoring data, which solves the problem that multimodal signals are difficult to fuse and analyze due to differences in noise characteristics, sampling frequency, and dimensions. This provides a high-quality, consistent, and standardized data foundation for subsequent high-precision, high-sensitivity dynamic feature mining and risk identification.

[0057] S3: Based on standardized multi-source functional monitoring data, extract the variation amplitude, complexity and stability characteristics of each monitoring signal, and quantify the functional degradation trend indicators.

[0058] The amplitude of change refers to the degree or rate of change of a signal over a period of time. It is not a single-point value, but a measure of the degree of change.

[0059] Complexity refers to the degree of disorder, irregularity, or unpredictability of a signal sequence's fluctuations. A highly complex signal may contain many irregular fluctuations, while a simple signal may be very smooth or have regular periodicity.

[0060] Among them, stability refers to the signal's ability to maintain its state or trend and resist fluctuations over a period of time. A stable signal has small fluctuations and a gradual change in its mean.

[0061] In one possible implementation, S3 specifically includes sub-steps S301 to S309:

[0062] S301: Obtain standardized multi-source functional monitoring data through a sliding time window. For each monitoring signal, calculate the difference between the multi-source functional monitoring data of two adjacent sampling points, divide it by the sampling interval, and then take the absolute value to obtain the functional change rate amplitude value.

[0063] In this context, a sliding time window refers to a fixed-length time interval that slides continuously along a timeline. The analysis is not based on the entire historical data, but rather dynamically focuses on data from a recent period.

[0064] S302: The decay weight value is obtained by performing natural exponentiation on the negative of the time decay weight factor as an exponent.

[0065] Among them, the time decay weighting factor is a dynamically adjusted parameter used to control the importance of historical data in the current calculation.

[0066] S303: Multiply the magnitude of the rate of change of function by the attenuation weight value to obtain the time-weighted rate of change of function.

[0067] S304: Accumulate the time-weighted rate of change of function for all sampling points within the sliding time window, and divide by the length of the sliding time window to obtain the intensity value of the change in function.

[0068] S305: Calculate the mean and standard deviation of the multi-source functional monitoring data for each monitoring signal within a sliding time window.

[0069] S306: The sample entropy value is calculated by evaluating the fluctuation characteristics of each monitoring signal within the sliding time window based on the mean and standard deviation of the multi-source functional monitoring data, and the functional complexity amplitude is obtained by taking the square root of the sample entropy value.

[0070] Among them, sample entropy refers to an algorithm used to measure the complexity and irregularity of time series data. The higher the entropy value, the more complex and unpredictable the sequence is.

[0071] S307: Divide the sliding time window into equal-length sub-blocks, calculate the average value of the multi-source functional monitoring data of each sub-block, and square the difference between the average values ​​of adjacent sub-blocks, calculate the mean, and then take the square root to obtain the stability scale.

[0072] S308: The robustness correction factor is obtained by dividing the functional complexity magnitude by the sum of the stability scale and the minimum constant value.

[0073] S309: Multiply the functional change intensity value by the robustness correction factor to obtain the functional change rate gain value.

[0074] Specifically, a sliding time window is used, the length of which is dynamically adjusted based on the signal's periodicity, volatility, and sampling frequency. If the signal changes significantly and short-term fluctuations are obvious, the window length should be appropriately shortened. Conversely, if the signal changes smoothly, the window length can be extended to better capture the global trend. The preprocessed multi-source functional monitoring data sequence is acquired. For each monitoring signal, the difference between two adjacent sampling points is calculated, divided by the sampling interval, and the absolute value is taken to obtain the functional change rate amplitude value, accurately quantifying the rate of change and fluctuation range of the monitoring signal. The inverse of the time decay weight factor is used as the exponent for natural exponential operation to obtain the decay weight value. The time decay weight factor is adjusted according to the signal's temporal distribution characteristics to better reflect the signal's changing trend in the time series. The functional change rate amplitude value is multiplied by the decay weight value to obtain the time-weighted functional change rate, emphasizing the importance of the most recent data points while appropriately attenuating older data points. The time-weighted functional change rates of all sampling points within the window are accumulated and divided by the sliding time window length to obtain the functional change intensity value, comprehensively measuring the overall change intensity of the signal within the sliding window. Simultaneously, the mean and standard deviation of the multi-source functional monitoring data for each monitoring signal are calculated within the window. The fluctuation characteristics of the monitoring signal within the window are evaluated based on the mean and standard deviation of the multi-source functional monitoring data, capturing the fluctuation trend and stability of the signal. The sample entropy value is calculated, and the square root of the sample entropy value is used to obtain the functional complexity amplitude. The sample entropy value reflects the complexity of the signal, and the square root can provide a smoother complexity index.

[0075] Furthermore, the time window is divided into equal-length sub-blocks. The number of sub-blocks is determined by dividing the window length into multiple equal sub-blocks. The number of sub-blocks depends on the signal volatility and the required analytical precision. In practical applications, the number of sub-blocks needs to be large enough to capture the detailed features of the signal, but not so many that it leads to excessive computation. Adaptive adjustments can be made using cross-validation. The analysis effect under different sub-block numbers is evaluated based on different sub-block numbers to select the number of blocks that most accurately reflects signal changes. If the signal volatility is strong, the number of sub-blocks can be increased to capture changes more precisely. If the signal is relatively stable, the number of sub-blocks can be reduced to optimize computational efficiency. The average value of the multi-source functional monitoring data for each sub-block is calculated. The square difference between the average values ​​of adjacent sub-blocks is squared, and the average of the squared differences of the average values ​​of all adjacent sub-blocks is taken. The square root of the average value is then used to obtain a stability scale. The stability scale measures the volatility of the signal within adjacent time periods, helping to understand the amplitude of signal fluctuations on the time axis. The robustness correction factor is obtained by dividing the functional complexity magnitude by the sum of the stability scale and the minimum constant value. This factor further adjusts the calculation of the functional change rate, taking into account both complexity and stability factors. Multiplying the functional change intensity value by the robustness correction factor yields the functional change rate gain value, which integrates the signal's change intensity, complexity, and stability, and is a core quantitative indicator of functional degradation trends.

[0076] Optionally, the specific formula for the gain value of the rate of change of function is:

[0077]

[0078] in, It represents the gain value of the rate of functional change, which quantitatively characterizes the rate, direction and stability of an individual's function changes over time. It is a basic indicator for identifying potential functional degradation trends, and directly reflects whether functional changes continue to accelerate or tend to fluctuate abnormally. It is used to determine whether a certain physiological function has undergone latent degradation. Essentially, it is a trend measurement model of "time decay rate of change integral plus stability normalization". Indicates the length of the sliding time window. This represents the multi-source functional monitoring data of the nth monitoring signal, which represents a certain functional state or physiological signal of an individual. This represents a time variable, indicating any point in time within the sliding time window. The sample entropy value represents the complexity and randomness of the fluctuations in the monitored signal within the window; a larger value indicates more disordered signal fluctuations. This represents a stability scale, characterizing the difference in signal stability across different time segments; a smaller value indicates a more stable signal. Represents a very small constant value, taking values ​​of , The time decay weighting factor is used to extract the functional change rate amplitude sequence of each monitored signal within a sliding time window, and calculate the variance and autocorrelation coefficient. A comprehensive index is constructed based on the variance and autocorrelation coefficient, where the comprehensive index equals the ratio of the variance to itself plus a minimum constant, multiplied by the complement of the autocorrelation coefficient. This comprehensive index is then mapped to the time decay weighting factor through a monotonic mapping relationship. This mapping relationship can be implemented using linear or exponential mapping. When the signal fluctuates drastically, has a large variance, and low autocorrelation, the time decay weighting factor takes a larger value to accelerate the decay of historical sample weights. When the signal changes smoothly, has a small variance, and high autocorrelation, the time decay weighting factor takes a smaller value to extend the time memory range. The time decay weighting factor ranges from 0.01 to 0.5. This represents a robust correction factor used to improve the accuracy of signal processing. It smooths out signal fluctuations across different time windows, thereby enabling the overall functional change trend to more accurately reflect the true functional degradation and reducing the impact of outlier data on the final analysis results.

[0079] For example, consider the functional change intensity, sample entropy, stability scale, and functional change rate gain of a detection signal within different windows. Specifically, window 1 corresponds to a functional change intensity of 0.18, sample entropy of 0.55, stability scale of 0.16, and functional change rate gain of 0.834. Window 2 corresponds to a functional change intensity of 0.22, sample entropy of 0.62, stability scale of 0.18, and functional change rate gain of 0.962. Window 3 corresponds to a functional change intensity of 0.27, sample entropy of 0.74, stability scale of 0.21, and functional change rate gain of 1.106. Window 4 corresponds to a functional change intensity of 0.25, sample entropy of 0.69, stability scale of 0.20, and functional change rate gain of 1.038. Window 5 corresponds to a functional change intensity of 0.31, sample entropy of 0.81, stability scale of 0.23, and functional change rate gain of 1.213.

[0080] Reference manual attached Figure 2 The diagram shows a trend graph of the functional change rate gain value provided by an embodiment of the present invention.

[0081] Appendix Figure 2The horizontal axis represents the time window number, and the vertical axis represents the functional rate of change gain. The line graph shows the overall trend of the functional rate of change gain. It can be seen that as the time window increases, the functional rate of change gain gradually increases, indicating that the functional rate of change gradually increases over time, which may indicate that the rate of functional degradation is accelerating. However, the functional rate of change gain decreases from window 3 to window 4, indicating that the gain of the functional rate of change slows down during certain periods. This decrease may indicate that the rate of functional degradation has slowed down or entered a temporary stable state. The functional rate of change gain increases again in window 5, indicating that the functional rate of change resumes growth during this stage, which may mean that the functional degradation process accelerates again in subsequent periods. Overall, the functional rate of change gain increases with the window number, indicating that functional degradation is gradually intensifying. Although there are short-term fluctuations, the overall trend is still upward, indicating that the intensification of functional degradation is a long-term process and may be accompanied by periodic fluctuations.

[0082] In this embodiment of the invention, by integrating the variation amplitude, complexity and stability characteristics and introducing time decay weighting and robustness correction mechanisms, a functional degradation trend index that can reflect the long-term evolution trend of a signal with high fidelity is constructed. This solves the problems of existing methods being insensitive to capturing slow, continuous and potentially fluctuating functional degradation processes and being susceptible to noise interference, and achieves early, accurate and robust identification of potential progressive functional degradation.

[0083] S4: Based on standardized multi-source functional monitoring data, the sudden change characteristics of the monitoring signals are extracted, short-term fluctuations and energy anomalies are quantified, and the functional burst response intensity value is evaluated.

[0084] Among them, abrupt change characteristics refer to the pattern characteristics of a signal that deviate significantly from its recent normal state within a short period of time. It is not a slow change, but a rapid, non-linear transition or jump.

[0085] Short-term fluctuations refer to the rapid fluctuations and changes of signals on a time scale of seconds or less.

[0086] In one possible implementation, S4 specifically includes sub-steps S401 to S404:

[0087] S401: Within the same sliding time window, for each monitoring signal, the functional deviation value is obtained by dividing the difference between the current multi-source functional monitoring data and the mean of the multi-source functional monitoring data by the sum of the standard deviation of the multi-source functional monitoring data and the minimum constant value.

[0088] S402: Calculate the first-order and second-order differences for the multi-source functional monitoring data sequences of each monitoring signal, and obtain the parameter-free curvature rate by dividing the absolute value of the second-order difference by the sum of the absolute value of the first-order difference and the minimum constant value.

[0089] Among them, the nonparametric curvature rate is a dimensionless indicator that describes the degree of abrupt "bending" or "turning" of a signal trajectory. It does not concern itself with the absolute speed of change, but rather with whether the speed changes abruptly.

[0090] S403: Perform wavelet decomposition on the multi-source functional monitoring data corresponding to each monitoring signal to obtain the energy values ​​of the detail components and the corresponding static stable component energy quantiles at each scale. Divide the energy values ​​of the detail components at each scale by the sum of the corresponding static stable component energy quantiles and the minimum constant value, and take the square root of the result with the largest ratio at each scale to obtain the multi-scale energy ratio.

[0091] Wavelet decomposition is an advanced signal processing method that can simultaneously decompose a signal into different time and frequency scales. It is better at capturing the local time-frequency characteristics of a signal than the traditional Fourier transform.

[0092] The detail component energy value refers to the signal energy carried by the detail component at a specific scale. A high energy value indicates that the signal fluctuates violently within that frequency range.

[0093] Among them, the static steady component energy quantile usually refers to a certain quantile of the approximate component energy, representing the energy level of the stable, normal part of the signal.

[0094] S404: The functional mutation response intensity value is obtained by multiplying the functional deviation value, the sum of the constant and the non-parametric curvature rate, and the multi-scale energy ratio.

[0095] Specifically, within the same window, for each monitoring signal, the functional deviation value is obtained by dividing the difference between the current multi-source functional monitoring data and the mean of the multi-source functional monitoring data by the sum of the standard deviation and the minimum constant value of the multi-source functional monitoring data. The functional deviation value reflects the degree of deviation of the monitoring signal from the mean, which helps to capture abnormal changes in the signal. First-order and second-order differences are calculated for the multi-source functional monitoring data sequences of each monitoring signal. The absolute value of the second-order difference is divided by the sum of the absolute value of the first-order difference and the minimum constant value to obtain the parameter-free curvature rate. The parameter-free curvature rate reflects the abruptness of changes in the monitoring signal, helping to identify sudden changes in function. Wavelet decomposition is performed on the multi-source functional monitoring data corresponding to each monitoring signal to obtain the energy values ​​of the detail components and the corresponding statically stable component energy quantiles at each scale. The energy values ​​of the detail components at each scale are divided by the sum of the corresponding statically stable component energy quantiles and the minimum constant value, and the square root of the result with the largest ratio at each scale is taken to obtain the multi-scale energy ratio. The multi-scale energy ratio helps to analyze the energy fluctuations of the signal at different frequency scales, further reflecting the multidimensional characteristics of functional anomalies. The functional deviation value, the sum of the constant and the non-parametric curvature rate, and the multi-scale energy ratio are multiplied together to obtain the functional mutation response intensity value.

[0096] Furthermore, by multiplying the three features—functional deviation, non-parametric curvature rate, and multi-scale energy ratio—the degree of abnormal fluctuation, abrupt changes, and energy change patterns of the monitored signal can be comprehensively reflected. Functional deviation measures the degree of deviation of the signal from the mean, helping to capture abnormal functional changes. Non-parametric curvature rate reveals the abruptness of signal changes, helping to identify sudden changes. Multi-scale energy ratio analyzes the energy fluctuations of the signal at different frequency scales, further reflecting the multidimensional characteristics of functional anomalies. The combination of these three features characterizes functional anomalies from three physical aspects: state shift, acceleration of change, and spectral energy redistribution. The product form emphasizes the synergistic occurrence of these three anomaly features. Only when the functional state simultaneously exhibits significant shift, rapid change, and multi-scale energy anomalies does the intensity of the functional abrupt response increase significantly. This effectively distinguishes real sudden functional degradation events from random noise or isolated fluctuations, more accurately identifies sudden functional degradation events, effectively suppresses false alarms, and ensures the stability and reliability of monitoring results.

[0097] Optionally, the specific formula for the functional mutation response intensity value is as follows:

[0098]

[0099] in, It represents the intensity value of functional mutation response, which is used to quantitatively describe the response intensity of an individual's functional signal when it experiences sudden fluctuations or critical turning points during time changes. That is, the degree to which the signal deviates from the steady state and changes rapidly in a short period of time. The design is essentially a mutation response model that is a superposition of "steady-state shift, mutation acceleration and energy anomaly", which is used to identify real functional mutations rather than random fluctuations in a short time scale. This represents the multi-source functional monitoring data of the nth monitoring signal. This represents the average value of multi-source functional monitoring data, indicating the average level of the signal within the current time window, and is used to construct a reference baseline. This represents the standard deviation of the multi-source functional monitoring data, characterizing the intensity of signal fluctuation within the current window; a larger standard deviation indicates more pronounced fluctuations. The second-order difference represents the multi-source functional monitoring data, and the acceleration of the signal change is used to identify nonlinear abrupt changes. The first-order difference represents the multi-source functional monitoring data, and the rate of signal change represents the signal change rate. It represents the energy value of the detail component, indicating the energy intensity of the signal at the local detail layer, and reflects the amplitude of short-time abrupt changes. This represents the energy quantile value of the static steady component, which serves as the energy benchmark for the stationary portion of the signal and is used for normalization comparison. Represents a very small constant value, taking values ​​of .

[0100] In this embodiment of the invention, by collaboratively extracting and fusing mutation features from three dimensions—amplitude deviation, curvature variation, and multi-scale energy distribution—a functional burst response intensity index capable of quantifying short-term functional abnormalities with high specificity is constructed. This solves the problems of inaccurate identification and poor noise resistance of traditional threshold methods for burst functional events with nonlinear and multi-frequency characteristics, and achieves accurate capture and low false alarm warning of real mutation events that mark stage-specific functional crises.

[0101] S5: Identify potential functional degradation trends and phased abrupt events based on functional degradation trend indicators and functional sudden response intensity values.

[0102] In one possible implementation, S5 specifically includes sub-steps S501 to S504:

[0103] S501: Calculate the functional change rate gain value and functional change response intensity value of all monitored signals and record them in time series form.

[0104] S502: Calculate the average and standard deviation of the gain value sequence of the functional change rate for each monitoring signal.

[0105] S503: When the duration for which the functional change rate gain value is higher than the average functional change rate gain value exceeds the maximum permissible threshold, and the functional change rate gain value is higher than the sum of the corresponding average value and standard deviation, it is determined that the monitoring signal has an abnormal upward trend in functional change rate, and the monitoring signal is marked as a potential functional degradation trend.

[0106] S504: When the intensity value of the functional mutation response is higher than the mutation threshold, the monitoring signal is marked as a stage mutation event.

[0107] It should be noted that those skilled in the art can set the maximum allowable threshold and the mutation threshold according to actual needs, and this invention does not limit them.

[0108] Specifically, the functional rate of change gain and functional mutation response intensity values ​​of all monitored signals are calculated in real time and recorded in time series form. Statistical analysis is performed on the functional rate of change gain value sequences of each monitored signal, calculating the mean and standard deviation. When the duration of a functional rate of change gain value exceeding the average functional rate of change gain value exceeds the maximum permissible threshold, and the functional rate of change gain value is higher than the sum of the corresponding mean and standard deviation, the monitored signal is determined to have an abnormal upward trend in functional rate of change, and is marked as a potential functional degradation trend. When the functional mutation response intensity value exceeds the mutation threshold, it is marked as a staged mutation event, and the sampling frequency is increased to ensure timely monitoring of functional degradation.

[0109] In this embodiment of the invention, personalized dynamic statistical baselines and duration determination rules based on time series are used to automatically make decisions on deeply quantified functional trends and mutation indicators. This transforms continuous data streams into clear structured event markers, overcoming the technical limitations of traditional methods that rely on subjective experience or fixed thresholds and cannot accurately distinguish between long-term gradual degradation and short-term accidental fluctuations. It achieves accurate and automatic identification of potential functional degradation trends and staged mutation events, providing clear and reliable input for subsequent risk classification and collaborative analysis, and improving the system's decision automation level.

[0110] S6: Based on the potential functional degradation trend and stage-specific abrupt events, extract the synergistic change characteristics of the monitoring signals and analyze the multidimensional functional coupling relationship.

[0111] Among them, the co-change characteristic refers to the consistency or correlation pattern of the changes (trends or abrupt changes) of multiple different monitoring signals during the time evolution process.

[0112] Among them, multidimensional functional coupling relationship refers to the strength and pattern of mutual connection and influence between different physiological functional dimensions.

[0113] In one possible implementation, S6 specifically includes sub-steps S601 to S604:

[0114] S601: Using the functional change rate gain value and functional mutation response intensity value calculated from each monitoring signal as input variables, construct the functional change rate matrix and the mutation response intensity matrix respectively.

[0115] S602: Calculate the Pearson correlation coefficient matrix for the functional change rate matrix to evaluate the synchronicity of the change trends among different monitoring signals.

[0116] The Pearson correlation coefficient is an indicator that measures the degree to which two variables are "in sync" in their trends of change.

[0117] S603: Calculate the covariance matrix of the sudden change response intensity matrix, and use multidimensional scaling analysis to reduce the dimensionality of the covariance matrix, generating a thermodynamic matrix reflecting the coordinated changes of the monitoring signals. Visualize the thermodynamic matrix to generate feature maps reflecting the coordinated fluctuation intensity of different monitoring signals.

[0118] Among them, multidimensional scaling analysis refers to a dimensionality reduction and visualization technique. It maps high-dimensional data to a low-dimensional space based on the similarity or difference matrix between research objects, while preserving the original distance relationships between objects as much as possible.

[0119] Among them, the intensity of co-oscillation refers to the degree of co-occurrence of abnormal fluctuations in different signals during a mutation event, as shown by MDS analysis and heatmap visualization.

[0120] S604: Normalize the calculation results of the functional change rate matrix and the mutation response intensity matrix to construct a multidimensional functional correlation feature set, and use the multidimensional functional correlation feature set to extract the cross-signal cooperative change pattern, generate and output the feature map reflecting the degree of coupling of multidimensional physiological functional states.

[0121] Among them, the collaborative change pattern refers to the representative joint change template across multiple signals extracted by data mining such as clustering and principal component analysis on a multidimensional functional correlation feature set (composed of normalized trend and mutation matrices).

[0122] Among them, the degree of coupling of multidimensional physiological functional states refers to a quantitative and comprehensive measure used to characterize the strength of the overall linkage and coordination between multiple physiological functional systems of an individual within a specific time window.

[0123] Specifically, the functional change rate gain value, functional mutation response intensity value, corresponding monitoring signal label, and multidimensional functional correlation feature set of all monitoring signals are written into the disability assessment risk data mining database.

[0124] Specifically, the functional change rate gain value and functional mutation response intensity value calculated from each monitoring signal are used as input variables. The functional change rate gain value or functional mutation response intensity value of the monitoring signal are listed in the two-dimensional dimension of time and signal, with the row as the time window, and the functional change rate matrix and mutation response intensity matrix are constructed respectively.

[0125] For example, during monitoring, if the rates of change of acceleration and grip force signals are large, it may indicate a significant change in the individual's motion state, and the corresponding rate of change gain value will be reflected in this case. A Pearson correlation coefficient matrix is ​​calculated from the functional rate of change matrix to assess the synchronicity of the changing trends among different monitoring signals. For instance, when acceleration and heart rate signals change synchronously within the same time period, it may indicate a very close relationship between the individual's movement and heart rate during a particular activity. A covariance matrix is ​​calculated from the abrupt response intensity matrix, and multidimensional scaling analysis is used to reduce the dimensionality of the covariance matrix, generating a thermodynamic matrix reflecting the coordinated changes in the monitoring signals. The thermodynamic matrix is ​​then visualized to generate a feature map reflecting the coordinated fluctuation intensity of different monitoring signals.

[0126] For example, abrupt changes in grasping force and electromyographic RMS signals may reflect abnormal fluctuations in an individual's grasping movements. Covariance mapping can help quantify the intensity of these fluctuations. The calculation results of the functional change rate matrix and the mutation response intensity matrix are normalized to ensure that different signal data can be compared under the same standard. A multidimensional functional correlation feature set is constructed, and this feature set is used to further extract cross-signal cooperative change patterns, generating and outputting a feature map reflecting the coupling degree of multidimensional physiological functional states. The multidimensional functional correlation feature set is formed by concatenating the corresponding column vectors of the functional change rate matrix and the mutation response intensity matrix within the same time window, used to characterize the joint features of each monitored signal in both trend change and sudden response dimensions. Correlation analysis and dimensionality reduction projection processing are performed on the multidimensional functional correlation feature set to extract synchronous change patterns and cooperative fluctuation structures of different monitored signals during temporal evolution. The feature map is generated by visually mapping the dimensionality-reduced cooperative features according to signal dimension and intensity, used to intuitively reflect the coupling degree and change distribution between multidimensional physiological functional states.

[0127] For example, when cadence and heart rate signals exhibit synchronous fluctuations within the same time window, this phenomenon indicates a certain degree of functional coupling between cadence and heart rate during a particular exercise or activity. This synchronicity reflects dynamic changes in an individual's physiological state, especially during tasks such as gait training or aerobic exercise, where cadence and heart rate fluctuations are typically closely correlated. The generated feature map can then identify strong functional coupling, further providing information on the individual's functional decline or recovery. The functional change rate gain values, functional mutation response intensity values, corresponding monitoring signal labels, and multidimensional functional correlation feature sets of all monitored signals are written into a disability risk data mining database to support subsequent big data analysis and risk assessment.

[0128] In this embodiment of the invention, by integrating trend correlation analysis and mutation co-mapping, and combining it with dimensionality reduction visualization technology, the co-change patterns and coupling relationships of multiple signals are systematically extracted and intuitively displayed from high-order features of multiple signals. This solves the technical limitations of existing assessment methods that view each physiological indicator in isolation and cannot reveal the intrinsic relationship of systemic dysfunction. It realizes a deep characterization and visual insight into the co-working state of an individual's multidimensional physiological function system, providing a key basis for accurately identifying systemic risks.

[0129] S7: Based on the characteristics of coordinated change, the trend of functional degradation and the intensity of sudden functional response, construct basic quantitative parameters for risk assessment, and combine them with multidimensional functional coupling relationships to quantify the degree of multidimensional functional degradation of individuals.

[0130] In one possible implementation, S7 specifically includes sub-steps S701 to S707:

[0131] S701: Obtain the functional change rate gain value and functional change response intensity value of all monitored signals, and receive the Pearson correlation coefficient matrix of the functional change rate matrix. Take the absolute value of the correlation coefficient between all monitored signals and calculate the average value as the overall coordinated change intensity value.

[0132] S702: By using a sliding time window, the average energy intensity of the monitoring signal is obtained by squaring the changes in the multi-source functional monitoring data at adjacent times of each monitoring signal and averaging them. The average energy intensity of all monitoring signals is accumulated and averaged to obtain the overall average energy value of the current time window. The overall average energy value within the historical preset value window is obtained, and the median is selected to obtain the reference energy value. The average energy factor is obtained by dividing the overall average energy value of the current time window by the reference energy value.

[0133] It should be noted that those skilled in the art can set the size of the historical preset value window according to actual needs, and this invention does not limit this.

[0134] S703: Within the sliding time window, the ratio of the standard deviation of the multi-source functional monitoring data of each monitoring signal to the mean of the multi-source functional monitoring data is used as the signal robustness value to complete the construction of basic quantization parameters.

[0135] Specifically, the system acquires the functional change rate gain and functional mutation response intensity values ​​of all monitored signals, and receives the Pearson correlation coefficient matrix of the functional change rate matrix. The absolute values ​​of the correlation coefficients between all monitored signals in the Pearson correlation coefficient matrix are taken and averaged to obtain the overall coordinated change intensity value, which helps to comprehensively evaluate the coordinated changes between different signals. Based on a sliding time window, the average energy intensity of the monitored signal is obtained by squaring the changes in multi-source functional monitoring data at adjacent times for each monitored signal and averaging the results. This effectively captures the instantaneous fluctuation energy of the signal. The average energy intensities of all monitored signals are accumulated and averaged to obtain the overall average energy value for the current time window. The overall average energy values ​​over N historical windows are obtained, and the median is selected to obtain the reference energy value. The selection of N considers the balance between data timeliness and noise suppression, with a preferred range of 5 to 20 windows. The average energy factor is obtained by dividing the overall average energy value of the current time window by the reference energy value, which can quantify the change in signal activity. Within the sliding time window, the mean and standard deviation of the multi-source functional monitoring data for each monitoring signal are obtained, and the ratio of the standard deviation to the mean is used as the signal robustness value to effectively assess the stability and volatility of the monitoring signal and reflect whether there are abnormal fluctuations in the signal.

[0136] S704: Multiply the functional change rate gain value and functional mutation response intensity value corresponding to each monitoring signal, and divide by the sum of a constant and the signal robustness value to obtain the risk component ratio.

[0137] S705: The functional coupling effect value is obtained by summing the risk component ratios of each monitoring signal and averaging them.

[0138] S706: The cooperative change correction ratio is obtained by dividing the overall cooperative change intensity value by the average energy factor.

[0139] S707: Multiply the functional coupling effect value by the synergistic change correction ratio, take the opposite number as the exponent for natural exponent calculation, and subtract the result of the natural exponent calculation from the constant to obtain the functional degradation risk response value, thus completing the quantification of the degree of multidimensional functional degradation of an individual.

[0140] Specifically, the risk component ratio is obtained by multiplying the functional change rate gain and functional mutation response intensity value corresponding to each monitoring signal, and then dividing by the sum of a constant and the signal robustness value. This risk component ratio, by combining the change rate and the mutation response intensity, reflects the comprehensive change of the signal. The risk component ratios of each monitoring signal are accumulated and averaged to obtain the functional coupling effect value, quantifying the impact of coordinated changes between different signals on functional degradation. The overall coordinated change intensity value is divided by the average energy factor to obtain the coordinated change correction ratio, adjusting the functional degradation trend according to the energy intensity and emphasizing the comprehensive fluctuation effect of the signal. The functional coupling effect value is multiplied by the coordinated change correction ratio, and the negative number is used as the exponent for natural exponential calculation. The result of the natural exponential calculation is subtracted from the constant to obtain the functional degradation risk response value. This value combines the coordinated changes and correction ratio of the signal, ultimately quantifying the individual functional degradation risk. The exponential calculation maps the cumulative effect of coordinated degradation of multiple signals to a bounded, monotonically increasing risk response value, avoiding the infinite amplification of risk due to linear superposition, while highlighting the nonlinear characteristics of approaching the degradation state from a stable state.

[0141] It should be noted that the functional degradation risk response value maps the degradation trend, sudden response, and cross-signal synergistic change relationship of multi-source functional monitoring signals into a unified risk response indicator. Its practical significance lies in the comprehensive risk quantification and state discrimination level. By first characterizing the risk contribution of a single signal when continuous changes and sudden disturbances coexist, then integrating the synergistic effect of multiple signals and introducing energy intensity correction, and finally constraining the risk output range in an exponential form, it can effectively distinguish between local fluctuations and overall functional degradation, avoiding misjudgments caused by the linear superposition of risks.

[0142] Optionally, the specific formula for the functional degradation risk response value is as follows:

[0143]

[0144] in, It represents the risk response value of functional degradation, which is used to comprehensively quantify the overall functional degradation risk of an individual under multi-source physiological function monitoring. It represents the overall functional degradation intensity at the current moment. Essentially, it is a risk function that combines "multi-signal synergistic degradation accumulation, energy normalization correction and exponential risk mapping" to robustly assess the overall functional degradation risk while avoiding linear superposition amplification. This indicates the number of monitored signals. This value represents the overall coordinated change intensity, characterizing the synchronicity of changes among multiple monitoring signals. A larger value indicates that multiple signals fluctuate or deteriorate simultaneously. The average energy factor represents the overall active energy level of the monitoring signal within the current time window; a higher value indicates more abundant individual physiological activity. It represents the gain value of the rate of change of function, describes the rate and trend of change of the nth functional signal over time, and reflects the speed of functional degradation. This represents the intensity value of the functional mutation response, indicating the intensity of a short-term mutation or sharp fluctuation in the nth functional signal, used to identify sudden anomalies. This represents the signal robustness value, which measures the stability of the signal within a time window. It is the ratio of the standard deviation to the mean; the larger the value, the more volatile the signal. It represents the risk component ratio, reflecting the combined effect of the rate of change, mutation intensity, and stability correction of a single physiological function, and reflects the instantaneous risk response value at the individual functional level. This represents the synergistic change correction ratio, indicating the amplification or suppression effect of energy-synergistic relationships on risk at the group level.

[0145] For example, the functional rate of change gain, functional mutation response intensity, signal robustness, and risk component ratio of five monitoring signals within the same window are analyzed. Specifically, the functional rate of change gain for the monitoring signal acceleration is 0.42, functional mutation response intensity is 0.65, signal robustness is 0.21, and risk component ratio is 0.226. The functional rate of change gain for the monitoring signal EMG RMS is 0.36, functional mutation response intensity is 0.58, signal robustness is 0.18, and risk component ratio is 0.177. The functional rate of change gain for the monitoring signal heart rate is 0.51, functional mutation response intensity is 0.72, signal robustness is 0.24, and risk component ratio is 0.296. The functional rate of change gain for the monitoring signal grip force is 0.47, functional mutation response intensity is 0.68, signal robustness is 0.19, and risk component ratio is 0.268. The monitored skin conductance showed a functional rate of change gain of 0.39, a functional mutation response intensity of 0.63, a signal robustness of 0.20, and a risk component ratio of 0.204. With the overall synergistic change intensity set at 0.85 and the average energy factor at 1.12, the functional coupling effect was calculated to be 0.2342, and the functional degradation risk response was 0.162, based on the data table.

[0146] Reference manual attached Figure 3 The diagram illustrates a radar schematic showing the risk ratio of various monitoring signals according to an embodiment of the present invention.

[0147] Appendix Figure 3The broken line represents the distribution of risk component ratios for different monitoring signals, and the dashed circle represents the functional coupling effect value calculated by the system. The background area in the figure is divided into safe zone, concern zone, and risk zone according to risk level, each marked with different labels. It can be seen that the risk component ratios for different monitoring signals differ. The risk component ratios for central rate and grip strength signals are 0.296 and 0.268, respectively, both higher than the functional coupling effect value, indicating that these two types of signals have larger functional fluctuations and sudden response intensity within the current time window, contributing significantly to the overall functional decline risk. The risk component ratios for acceleration and skin conductance signals are 0.226 and 0.204, respectively, close to or slightly lower than the functional coupling effect value, indicating that the functional state is in a relatively stable stage. The risk component ratio for electromyography (EMG) RMS signals is the lowest at 0.177, indicating smaller fluctuations and higher stability in muscle contraction intensity. Furthermore, the current individual's functional decline risk response value is generally at the edge of the concern zone, with central rate and grip strength signals being the main risk drivers, and should be the focus of monitoring in subsequent assessments and interventions. Other signals fluctuated relatively smoothly, and the overall functional status was still stable, but short-term trends need to be monitored to prevent sudden degradation.

[0148] In this embodiment of the invention, by constructing a multi-level parameter system that integrates independent risk of the signal, system synergy, and overall physiological state background, and by adopting a nonlinear risk convergence model, the dispersed multi-dimensional feature information is organically combined into a comprehensive quantitative index of individual functional degradation risk. This solves the problem that the traditional linear weighted model cannot characterize the nonlinear accumulation of risk and the emergence characteristics of systemic imbalance. It achieves accurate and dynamic quantification of comprehensive and nonlinear growth risk caused by the synergistic deterioration of multiple system functions, and provides a core decision-making basis for intelligent hierarchical early warning.

[0149] S8: Based on the degree of multidimensional functional decline of an individual, conduct dynamic identification and hierarchical management of disability risk assessment.

[0150] In one possible implementation, S8 specifically refers to:

[0151] By using a sliding time window, the functional degradation risk response value of the monitoring signal is calculated in real time, and the functional degradation risk response value is compared with the first risk threshold and the second risk threshold.

[0152] When the risk response value of functional degradation is less than or equal to the first risk threshold, the monitoring signal is determined to be in a stable state, and the normal sampling frequency and data monitoring are maintained.

[0153] When the risk response value of functional degradation is greater than the first risk threshold and less than or equal to the second risk threshold, it is determined to be a potential risk state, the corresponding monitoring signal is marked, the early warning signal is triggered and promoted, the sampling frequency is increased, and the monitoring signal with the risk component ratio greater than the intensity threshold is retested.

[0154] When the functional degradation risk response value exceeds the second risk threshold, a risk state is identified. The Pearson correlation coefficient matrix of each monitoring signal within the corresponding sliding time window is obtained, and it is determined whether the proportion of multiple signals changing in the same direction exceeds the proportion threshold. If so, the risk state is confirmed as true degradation. Otherwise, the multi-source functional monitoring data and the functional degradation risk response value are verified.

[0155] For monitoring signals where the risk component ratio is greater than the intensity threshold, a key monitoring instruction is generated, the sampling frequency is increased and the sliding time window is shortened, and a function degradation intervention prompt is pushed out.

[0156] Specifically, the functional degradation risk response value and the corresponding grading results are written into the disability assessment risk data mining database.

[0157] Specifically, the functional degradation risk response value is calculated in real time based on a sliding time window, and the functional degradation risk response value is... Compare with multi-level risk thresholds R1 and R2. When When R1 ≤ R1, the system is considered to be in a stable state. The normal sampling frequency and data monitoring are maintained to ensure basic monitoring can continue without triggering unnecessary interventions, thus ensuring long-term stability. When R1 < When R² is ≤2, it is determined to be a potential risk state, the corresponding monitoring signal is marked, an early warning signal is triggered and promoted, the sampling frequency is increased, preferably to 1.5 to 2 times the original frequency, to enhance the sensitivity and timeliness of the data, and monitoring signals with risk component ratios greater than the intensity threshold are retested. When R² > 2, a risk status is identified, indicating a significant increase in the individual's risk of functional decline. Cross-validation is then performed: the Pearson correlation coefficient matrix of each monitoring signal within the corresponding time window is obtained. If the proportion of multiple signals changing in the same direction is greater than the proportion threshold, the risk status is confirmed as true decline; otherwise, multi-source functional monitoring data and functional decline risk response values ​​are reviewed. For monitoring signals with risk component ratios greater than the intensity threshold, a key monitoring instruction is generated, increasing the sampling frequency and shortening the sliding time window. Ideally, the sampling frequency can be increased to 2 to 3 times the original frequency, and the sliding window length should be shortened to 30% to 50% of the original length to ensure real-time tracking of individual physiological changes. Simultaneously, functional decline intervention prompts are pushed out. The functional decline risk response value and the corresponding grading results are written into the disability assessment risk data mining database.

[0158] It should be noted that those skilled in the art can set the values ​​of the first risk threshold, the second risk threshold, the proportion threshold, and the intensity threshold according to actual needs, and this invention does not limit these settings.

[0159] In this embodiment of the invention, multi-level threshold dynamic judgment and adaptive monitoring strategy adjustment are realized based on comprehensive risk quantification indicators. Combined with a collaborative verification mechanism, a closed-loop response process from risk identification and cross-validation to precise intervention is constructed. This solves the technical problems of rigid response, high false alarm rate and inability to dynamically optimize monitoring resources according to risk level in traditional early warning systems. It realizes intelligent hierarchical management and precise and efficient prevention and control of disability assessment risks.

[0160] S9: Based on the grading results, by continuously monitoring and visualizing functional degradation trend indicators, functional emergency response intensity values, and individual multidimensional functional degradation levels, the system dynamically displays functional degradation trends and emergencies. Furthermore, it optimizes thresholds based on historical data to achieve closed-loop self-learning of prevention and control strategies.

[0161] In one possible implementation, S9 specifically refers to:

[0162] Continuously monitor the functional change rate gain value, functional mutation response intensity value, and functional degradation risk response value. Through time series, visualize the dynamic change trend of multidimensional functions over time, the distribution of sudden response intensity, and the temporal evolution law of risk intensity.

[0163] Among them, the temporal evolution law refers to the dynamic change process and trend of the "functional degradation risk response value" and its driving factors over time, exhibiting a specific pattern. It focuses on how the risk changes, rather than just the magnitude of the risk at a certain moment.

[0164] It displays the markers of potential functional decline trends and stage-specific abrupt events, and outputs the risk component ratios of each monitoring signal in descending order. It also presents a feature map of the coupling degree of multidimensional physiological functional states.

[0165] When the function degradation risk response value of a continuously preset numerical window is detected to be rising, a trend warning layer is generated, and the inflection point of change is highlighted with a dynamic identifier.

[0166] It should be noted that those skilled in the art can set the size of the continuous preset value window according to actual needs, and this invention does not limit this.

[0167] Based on historical functional change rate gain values, functional mutation response intensity values, and corresponding functional degradation risk response values, a random forest regression algorithm is used to dynamically adjust mutation thresholds and multi-level risk thresholds. Combined with visualization results and functional degradation risk response value grading results, a functional degradation risk assessment report is generated.

[0168] It should be noted that those skilled in the art can set the size of historical data optimization threshold, mutation threshold, and multi-level risk threshold according to actual needs, and this invention does not limit these settings.

[0169] Specifically, the system continuously monitors the functional change rate gain, functional mutation response intensity, and functional degradation risk response value. Time-series visualization is used to display the dynamic trends of multidimensional functional changes over time, the distribution of sudden response intensity, and the temporal evolution of risk intensity, ensuring the timeliness and real-time nature of the monitoring data. Potential functional degradation trends and staged mutation event markers are simultaneously displayed, and the risk component ratios of each monitored signal are output in descending order to facilitate real-time assessment of the risk priority of the monitored signals. A feature map of the coupling degree of multidimensional physiological functional states is also displayed, reflecting the collaborative change patterns among multiple signals and providing intuitive physiological state analysis. When a continuous increase in the functional degradation risk response value is detected for w consecutive windows, a trend warning layer is generated, and the inflection point is dynamically highlighted to reflect the trend change of potential risks in a timely manner. The value of w can be adjusted according to the specific application scenario and the characteristics of the monitored object. The optimal range for w is between 3 and 10 windows, which helps to eliminate short-term noise interference and ensure the stability and accuracy of the warning.

[0170] Furthermore, a feature matrix containing functional change rate gain, functional mutation response intensity, and functional degradation risk response value is constructed using historical data and used as input data for training. The random forest regression model generates multiple decision trees through multiple iterations, using the output of each tree to predict threshold adjustments. During training, the error between the functional degradation risk response value and a predefined risk threshold is used as the optimization objective to minimize the prediction error and improve model accuracy. After each iteration, the mutation threshold and multi-level risk thresholds are dynamically updated based on the prediction results, and the thresholds are periodically adjusted according to changes in actual monitoring data to ensure real-time response to individual physiological changes and automatic adaptation to different risk states and environmental changes. During threshold adjustment, the update frequency can be optimally set to once every 5 windows, depending on the specific application scenario, to ensure a balance between adaptive threshold adjustment and overall stability. Simultaneously, by combining the visualization results with the functional degradation risk response value classification results, a functional degradation risk assessment report is generated, including: analysis results of functional change rate gain value, functional mutation response intensity value and functional degradation risk response value, risk status classification results and corresponding intervention measures, visualization map of coordinated changes in monitoring signals and functional coupling degree, thresholds dynamically adjusted based on historical data, and optimization suggestions for prevention and control strategies.

[0171] In this embodiment of the invention, by deeply integrating multi-dimensional risk dynamic visualization with a machine learning-based historical data self-learning mechanism, a full-link intelligent monitoring system is constructed, from real-time status perception and trend warning to threshold adaptive optimization. This solves the limitations of traditional monitoring systems, such as unintuitive human-computer interaction, fixed warning strategies, and inability to self-improve with the evolution of individual status. It realizes closed-loop self-learning and continuous dynamic optimization of disability assessment risk prevention and control strategies, significantly improving the system's intelligence level and long-term practical effectiveness.

[0172] In practical applications, the system overcomes the technical bottleneck of multimodal signal fusion by real-time acquisition and preprocessing of multi-source functional monitoring data. Based on this, the system simultaneously extracts functional change rate gain values ​​reflecting long-term degradation trends and functional mutation response intensity values ​​characterizing short-term mutations from standardized data, achieving coordinated capture of both "slow variables" and "fast variables" of functional status. Furthermore, by analyzing the coordinated changes and coupling relationships among multiple signals, a functional degradation risk response value integrating independent risk, system synergy, and physiological background is constructed, and dynamic risk classification and adaptive monitoring are achieved based on this value. Finally, by combining multi-dimensional visualization and a threshold self-learning mechanism based on random forests, a complete closed loop of "monitoring-analysis-early warning-optimization" is formed. This not only significantly improves the sensitivity and accuracy of identifying gradual degradation and sudden events, but also, through intelligent closed-loop management, achieves a fundamental improvement in disability assessment risk prevention and control, moving from passive response to proactive prediction and from static thresholds to dynamic adaptation.

[0173] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0174] This invention, by simultaneously extracting the variation amplitude, complexity, and stability characteristics of multi-source functional monitoring signals, fully mines the deep degradation trend information contained in the data stream. Furthermore, it introduces a functional burst response intensity quantification method to dynamically identify short-term fluctuations and energy anomalies, improving the ability to characterize the dynamic evolution of functional states. This allows for the simultaneous capture of slow, continuous functional degradation trends and sudden, drastic, phased functional mutation events, enhancing the timeliness and accuracy of disability risk identification and reducing early warning delays and missed event reports. It also increases the system's ability to achieve early, accurate, and categorized early warnings.

[0175] Reference manual attached Figure 4 The diagram shows a structural schematic of a data mining-based disability risk identification and prevention analysis system provided by the present invention.

[0176] This invention also provides a data mining-based disability assessment risk identification and prevention analysis system 20, applied to the aforementioned data mining-based disability assessment risk identification and prevention analysis method, comprising:

[0177] Processor 201.

[0178] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201, they implement the data mining-based disability assessment risk identification and prevention analysis method as described in the method embodiment.

[0179] The data mining-based disability assessment risk identification and prevention analysis system 20 provided by this invention can execute the above-mentioned data mining-based disability assessment risk identification and prevention analysis method and achieve the same or similar technical effects. To avoid duplication, this invention will not elaborate further.

[0180] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0181] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0182] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0183] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0184] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0185] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0186] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0187] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0188] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0189] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0190] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0191] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0192] This invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the data mining-based disability assessment risk identification and prevention analysis method as described in the method embodiment.

[0193] The present invention provides a computer-readable storage medium that can implement the steps and effects of the data mining-based disability risk identification and prevention analysis method of the above-described method embodiments. To avoid repetition, the present invention will not elaborate further.

[0194] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0195] The following points need to be explained:

[0196] (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0197] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0198] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0199] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A data mining-based method for disability assessment risk identification and prevention analysis, characterized in that, include: S1: Collect multi-source functional monitoring data corresponding to each monitoring signal; S2: The multi-source functional monitoring data is preprocessed by filtering, resampling and normalization to obtain standardized multi-source functional monitoring data; S3: Based on the standardized multi-source functional monitoring data, extract the variation amplitude, complexity, and stability characteristics of each monitoring signal, and quantify the functional degradation trend index; S4: Based on the standardized multi-source functional monitoring data, the sudden change characteristics of the monitoring signals are extracted, short-term fluctuations and energy anomalies are quantified, and the functional burst response intensity value is evaluated. S5: Identify potential functional degradation trends and stage-specific sudden events based on the functional degradation trend index and the functional sudden response intensity value; S6: Based on the potential functional degradation trend and the staged abrupt change events, extract the synergistic change characteristics of the monitoring signals and analyze the multidimensional functional coupling relationship; S7: Based on the aforementioned collaborative change characteristics, the aforementioned functional degradation trend, and the aforementioned functional sudden response intensity, construct basic quantitative parameters for risk assessment, and combine them with the aforementioned multidimensional functional coupling relationship to quantify the degree of individual multidimensional functional degradation; S8: Based on the degree of multidimensional functional decline of the individual, conduct dynamic identification and hierarchical management of disability risk; S9: Based on the grading results, by continuously monitoring and visualizing the functional degradation trend indicators, the functional sudden response intensity values, and the individual multidimensional functional degradation degree, the functional degradation trend and sudden events are dynamically displayed; and the threshold is optimized based on historical data to achieve closed-loop self-learning of the prevention and control strategy.

2. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, S2 specifically includes: The multi-source functional monitoring data are aligned and resampled using a unified timestamp; Random noise suppression is performed on the acceleration and electromyography (RMS) data in the multi-source functional monitoring data by combining bandpass filtering and moving average. The drift trend of heart rate, heart rate variability and skin conductance in the multi-source functional monitoring data is removed by using the Kalman filter algorithm. Z-score normalization and sliding window smoothing are performed on all the multi-source functional monitoring data to obtain the standardized multi-source functional monitoring data.

3. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, S3 specifically includes: S301: Obtain the standardized multi-source functional monitoring data through a sliding time window. For each monitoring signal, calculate the difference between the multi-source functional monitoring data of two adjacent sampling points, divide it by the sampling interval, and then take the absolute value to obtain the functional change rate amplitude value. S302: The decay weight value is obtained by performing natural exponentiation on the negative of the time decay weight factor as the exponent; S303: Multiply the magnitude of the function change rate by the attenuation weight value to obtain the time-weighted function change rate; S304: Accumulate the time-weighted rate of change of the function at all sampling points within the sliding time window, and divide by the length of the sliding time window to obtain the value of the intensity of the function change; S305: Calculate the mean and standard deviation of the multi-source functional monitoring data of each monitoring signal within the sliding time window; S306: Evaluate the fluctuation characteristics of each monitoring signal within the sliding time window based on the mean and standard deviation of the multi-source functional monitoring data, calculate the sample entropy value, and take the square root of the sample entropy value to obtain the functional complexity amplitude. S307: Divide the sliding time window into equal-length sub-blocks, calculate the average value of the multi-source functional monitoring data of each sub-block, and square the difference between the average values ​​of adjacent sub-blocks, calculate the mean, and then take the square root to obtain the stability scale. S308: Divide the aforementioned functional complexity magnitude by the sum of the stability scale and the minimum constant value to obtain the robustness correction factor; S309: Multiply the functional change intensity value by the robust correction factor to obtain the functional change rate gain value.

4. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, S4 specifically includes: S401: Within the same sliding time window, for each of the monitoring signals, the difference between the current multi-source functional monitoring data and the mean of the multi-source functional monitoring data is divided by the sum of the standard deviation of the multi-source functional monitoring data and the minimum constant value to obtain the functional deviation value. S402: Calculate the first-order difference and the second-order difference for the multi-source functional monitoring data sequence of each of the monitoring signals respectively, and obtain the parameterless curvature rate by dividing the absolute value of the second-order difference by the sum of the absolute value of the first-order difference and the minimum constant value. S403: Perform wavelet decomposition on the multi-source functional monitoring data corresponding to each of the monitoring signals to obtain the energy values ​​of the detail components and the corresponding static steady component energy quantiles at each scale. Divide the energy values ​​of the detail components at each scale by the sum of the corresponding static steady component energy quantiles and the minimum constant value, and take the square root of the result with the largest ratio at each scale to obtain the multi-scale energy ratio. S404: The functional deviation value, the sum of the constant and the parameterless curvature rate, and the multiscale energy ratio are multiplied together to obtain the functional mutation response intensity value.

5. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, S5 specifically includes: S501: Calculate the functional change rate gain value and the functional abrupt response intensity value of all the monitored signals, and record them in time series form; S502: Calculate the average and standard deviation of the gain value sequence of the functional change rate of each of the monitoring signals; S503: When the duration for which the functional change rate gain value is higher than the average functional change rate gain value exceeds the maximum permissible threshold, and the functional change rate gain value is higher than the sum of the corresponding average value and standard deviation, it is determined that the monitoring signal has an abnormal upward trend in functional change rate, and the monitoring signal is marked as the potential functional degradation trend. S504: When the functional mutation response intensity value is higher than the mutation threshold, the monitoring signal is marked as the stage mutation event.

6. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, S6 specifically includes: S601: Using the functional change rate gain value and the functional mutation response intensity value calculated from each of the monitoring signals as input variables, construct the functional change rate matrix and the mutation response intensity matrix respectively; S602: Calculate the Pearson correlation coefficient matrix for the functional change rate matrix to evaluate the synchronicity of the change trends among different monitoring signals; S603: Calculate the covariance matrix of the mutation response intensity matrix, and use the multidimensional scaling analysis method to reduce the dimensionality of the covariance matrix to generate a thermodynamic matrix that reflects the coordinated changes of the monitoring signals; and generate a feature map that reflects the coordinated fluctuation intensity of different monitoring signals by visualizing the thermodynamic matrix. S604: Normalize the calculation results of the functional change rate matrix and the mutation response intensity matrix to construct a multidimensional functional association feature set, and use the multidimensional functional association feature set to extract cross-signal cooperative change patterns, generate and output feature maps reflecting the coupling degree of multidimensional physiological functional states.

7. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, Specifically, S7 includes: S701: Obtain the functional change rate gain value and the functional mutation response intensity value of all the monitoring signals, and receive the Pearson correlation coefficient matrix of the functional change rate matrix, take the absolute value of the correlation coefficient between all the monitoring signals and calculate the average value as the overall cooperative change intensity value; S702: By using a sliding time window, the average energy intensity of the monitoring signal is obtained by squaring the changes in the multi-source functional monitoring data at adjacent times of each monitoring signal and averaging them. The average energy intensity of all the monitoring signals is accumulated and averaged to obtain the overall average energy value of the current time window. The overall average energy value within the historical preset value window is obtained, and the median is selected to obtain the reference energy value. The average energy factor is obtained by dividing the overall average energy value of the current time window by the reference energy value. S703: Within the sliding time window, the ratio of the standard deviation of the multi-source functional monitoring data of each monitoring signal to the mean of the multi-source functional monitoring data is used as the signal robustness value to complete the construction of the basic quantization parameters. S704: Multiply the functional change rate gain value and the functional mutation response intensity value corresponding to each of the monitoring signals, and divide by the sum of a constant and the signal robustness value to obtain the risk component ratio; S705: The risk component ratios of each of the monitoring signals are summed and averaged to obtain the functional coupling effect value; S706: Divide the overall coordinated change intensity value by the average energy factor to obtain the coordinated change correction ratio; S707: Multiply the functional coupling effect value with the synergistic change correction ratio, take the opposite number as the exponent for natural exponent calculation, and subtract the natural exponent calculation result from the constant to obtain the functional degradation risk response value, thus completing the quantification of the individual's multidimensional functional degradation degree.

8. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, Specifically, S8 is: The functional degradation risk response value of the monitoring signal is calculated in real time by using a sliding time window, and the functional degradation risk response value is compared with a first risk threshold and a second risk threshold. When the functional degradation risk response value is less than or equal to the first risk threshold, the monitoring signal is determined to be in a stable state, and the normal sampling frequency and data monitoring are maintained. When the functional degradation risk response value is greater than the first risk threshold and less than or equal to the second risk threshold, it is determined to be a potential risk state, the corresponding monitoring signal is marked, an early warning signal is triggered and promoted, the sampling frequency is increased, and the monitoring signal with the risk component ratio greater than the intensity threshold is retested; When the functional degradation risk response value is greater than the second risk threshold, it is determined to be a risk state. The Pearson correlation coefficient matrix of each monitoring signal within the corresponding sliding time window is obtained, and it is determined whether the proportion of multiple signals changing in the same direction is greater than the proportion threshold. If so, then confirm that the risk status is a true degradation; otherwise, verify the multi-source functional monitoring data against the functional degradation risk response value. For monitoring signals where the risk component ratio is greater than the intensity threshold, a key monitoring instruction is generated, the sampling frequency is increased and the sliding time window is shortened, and a function degradation intervention prompt is pushed.

9. The method for disability assessment risk identification and prevention analysis based on data mining according to claim 1, characterized in that, Specifically, S9 is: Continuously monitor the functional change rate gain value, the functional mutation response intensity value, and the functional degradation risk response value. Through time series, visualize the dynamic change trend of multidimensional functions over time, the distribution of sudden response intensity, and the temporal evolution law of risk intensity. The system displays the potential functional degradation trend and the markers of the stage-specific abrupt events, and outputs the risk component ratios of each of the monitoring signals in descending order; Feature maps showcasing the degree of coupling between multidimensional physiological functional states; When the function degradation risk response value of the continuously preset numerical window is detected to be rising, a trend warning layer is generated, and the inflection point of change is highlighted with a dynamic identifier. Based on the historical functional change rate gain value, the functional mutation response intensity value and the corresponding functional degradation risk response value, the mutation threshold and multi-level risk threshold are dynamically adjusted through the random forest regression algorithm. The system combines visualization results with functional degradation risk response value grading results to generate a functional degradation risk assessment report.

10. A data mining-based system for disability assessment risk identification and prevention analysis, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the data mining-based method for disability assessment risk identification and prevention analysis as described in any one of claims 1 to 9.

Images